Advances in Computational Intelligence

13th International Work Conference on

Artificial Neural Networks IWANN 2015
Palma de Mallorca Spain June 10 12
2015 Proceedings Part II 1st Edition
Ignacio Rojas
Visit to download the full and correct content document:
https://textbookfull.com/product/advances-in-computational-intelligence-13th-internati
onal-work-conference-on-artificial-neural-networks-iwann-2015-palma-de-mallorca-sp
ain-june-10-12-2015-proceedings-part-ii-1st-edition-ignacio-rojas/
More products digital (pdf, epub, mobi) instant
download maybe you interests ...

Advances in Computational Intelligence 13th

International Work Conference on Artificial Neural
Networks IWANN 2015 Palma de Mallorca Spain June 10 12
2015 Proceedings Part I 1st Edition Ignacio Rojas
https://textbookfull.com/product/advances-in-computational-
intelligence-13th-international-work-conference-on-artificial-
neural-networks-iwann-2015-palma-de-mallorca-spain-
june-10-12-2015-proceedings-part-i-1st-edition-ignacio-rojas/

Advances in Computational Intelligence 14th

International Work Conference on Artificial Neural
Networks IWANN 2017 Cadiz Spain June 14 16 2017
Proceedings Part I 1st Edition Ignacio Rojas
https://textbookfull.com/product/advances-in-computational-
intelligence-14th-international-work-conference-on-artificial-
neural-networks-iwann-2017-cadiz-spain-
june-14-16-2017-proceedings-part-i-1st-edition-ignacio-rojas/

Bioinspired Computation in Artificial Systems

International Work Conference on the Interplay Between
Natural and Artificial Computation IWINAC 2015 Elche
Spain June 1 5 2015 Proceedings Part II 1st Edition
José Manuel Ferrández Vicente
https://textbookfull.com/product/bioinspired-computation-in-
artificial-systems-international-work-conference-on-the-
interplay-between-natural-and-artificial-computation-
iwinac-2015-elche-spain-june-1-5-2015-proceedings-part-ii-1st-ed/

Artificial Intelligence in Education 17th International

Conference AIED 2015 Madrid Spain June 22 26 2015
Proceedings 1st Edition Cristina Conati

https://textbookfull.com/product/artificial-intelligence-in-
education-17th-international-conference-aied-2015-madrid-spain-
june-22-26-2015-proceedings-1st-edition-cristina-conati/
Bioinformatics and Biomedical Engineering 5th
International Work Conference IWBBIO 2017 Granada Spain
April 26 28 2017 Proceedings Part II 1st Edition
Ignacio Rojas
https://textbookfull.com/product/bioinformatics-and-biomedical-
engineering-5th-international-work-conference-
iwbbio-2017-granada-spain-april-26-28-2017-proceedings-part-
ii-1st-edition-ignacio-rojas/

Formal Concept Analysis 13th International Conference

ICFCA 2015 Nerja Spain June 23 26 2015 Proceedings 1st
Edition Jaume Baixeries

https://textbookfull.com/product/formal-concept-analysis-13th-
international-conference-icfca-2015-nerja-spain-
june-23-26-2015-proceedings-1st-edition-jaume-baixeries/

Advances in Computational Intelligence 19th Mexican

International Conference on Artificial Intelligence
MICAI 2020 Mexico City Mexico October 12 17 2020
Proceedings Part II Lourdes Martínez-Villaseñor
https://textbookfull.com/product/advances-in-computational-
intelligence-19th-mexican-international-conference-on-artificial-
intelligence-micai-2020-mexico-city-mexico-
october-12-17-2020-proceedings-part-ii-lourdes-martinez-
villasenor/
Artificial Intelligence and Soft Computing 14th
International Conference ICAISC 2015 Zakopane Poland
June 14 18 2015 Proceedings Part I 1st Edition Leszek
Rutkowski
https://textbookfull.com/product/artificial-intelligence-and-
soft-computing-14th-international-conference-
icaisc-2015-zakopane-poland-june-14-18-2015-proceedings-
part-i-1st-edition-leszek-rutkowski/

Current Approaches in Applied Artificial Intelligence

28th International Conference on Industrial Engineering
and Other Applications of Applied Intelligent Systems
IEA AIE 2015 Seoul South Korea June 10 12 2015
Proceedings 1st Edition Moonis Ali
https://textbookfull.com/product/current-approaches-in-applied-
artificial-intelligence-28th-international-conference-on-
industrial-engineering-and-other-applications-of-applied-
 Ignacio Rojas
Gonzalo Joya
Andreu Catala (Eds.)

Advances in
LNCS 9095

Computational
Intelligence
13th International Work-Conference
on Artificial Neural Networks, IWANN 2015
Palma de Mallorca, Spain, June 10–12, 2015
Proceedings, Part II

123
Lecture Notes in Computer Science 9095
Commenced Publication in 1973
Founding and Former Series Editors:
Gerhard Goos, Juris Hartmanis, and Jan van Leeuwen

Editorial Board
David Hutchison
Lancaster University, Lancaster, UK
Takeo Kanade
Carnegie Mellon University, Pittsburgh, PA, USA
Josef Kittler
University of Surrey, Guildford, UK
Jon M. Kleinberg
Cornell University, Ithaca, NY, USA
Friedemann Mattern
ETH Zürich, Zürich, Switzerland
John C. Mitchell
Stanford University, Stanford, CA, USA
Moni Naor
Weizmann Institute of Science, Rehovot, Israel
C. Pandu Rangan
Indian Institute of Technology, Madras, India
Bernhard Steffen
TU Dortmund University, Dortmund, Germany
Demetri Terzopoulos
University of California, Los Angeles, CA, USA
Doug Tygar
University of California, Berkeley, CA, USA
Gerhard Weikum
Max Planck Institute for Informatics, Saarbrücken, Germany
More information about this series at http://www.springer.com/series/7407
Ignacio Rojas · Gonzalo Joya
Andreu Catala (Eds.)

Advances in
Computational
Intelligence
13th International Work-Conference
on Artificial Neural Networks, IWANN 2015
Palma de Mallorca, Spain, June 10–12, 2015
Proceedings, Part II

ABC
Editors
Ignacio Rojas Andreu Catala
University of Granada Polytechnic University of Catalonia
Granada Vilanova i la Geltrú
Spain Spain

Gonzalo Joya
University of Malaga
Malaga
Spain

ISSN 0302-9743 ISSN 1611-3349 (electronic)

Lecture Notes in Computer Science
ISBN 978-3-319-19221-5 ISBN 978-3-319-19222-2 (eBook)
DOI 10.1007/978-3-319-19222-2

Library of Congress Control Number: 2015939427

LNCS Sublibrary: SL1 – Theoretical Computer Science and General Issues

Springer Cham Heidelberg New York Dordrecht London

c Springer International Publishing Switzerland 2015
This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the
material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation,
broadcasting, reproduction on microfilms or in any other physical way, and transmission or information stor-
age and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now
known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication
does not imply, even in the absence of a specific statement, that such names are exempt from the relevant
protective laws and regulations and therefore free for general use.
The publisher, the authors and the editors are safe to assume that the advice and information in this book
are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the
editors give a warranty, express or implied, with respect to the material contained herein or for any errors or
omissions that may have been made.

Printed on acid-free paper

Springer International Publishing AG Switzerland is part of Springer Science+Business Media

(www.springer.com)
 Preface

We are proud to present the set of final accepted papers for the 13th edition of the
IWANN conference “International Work-Conference on Artificial Neural Networks”
held in Palma de Mallorca (Spain) during June 10–12, 2015.
IWANN is a biennial conference that seeks to provide a discussion forum for sci-
entists, engineers, educators, and students about the latest ideas and realizations in the
foundations, theory, models, and applications of hybrid systems inspired on nature (neu-
ral networks, fuzzy logic, and evolutionary systems) as well as in emerging areas related
to the above items. As in previous editions of IWANN, it also aims to create a friendly
environment that could lead to the establishment of scientific collaborations and ex-
changes among attendees. The proceedings will include all the presented communica-
tions to the conference. It has also foreseen the publication of an extended version of
selected papers in a special issue of several specialized journals (such as Neurocomput-
ing, Soft Computing, and Neural Processing Letters).
Since the first edition in Granada (LNCS 540, 1991), the conference has evolved and
matured. The list of topics in the successive Call for Papers has also evolved, resulting
in the following list for the present edition:

1. Mathematical and theoretical methods in computational intelligence. Mathe-

matics for neural networks. RBF structures. Self-organizing networks and meth-
ods. Support vector machines and kernel methods. Fuzzy logic. Evolutionary and
genetic algorithms.
2. Neurocomputational formulations. Single-neuron modelling. Perceptual mod-
elling. System-level neural modelling. Spiking neurons. Models of biological learn-
ing.
3. Learning and adaptation. Adaptive systems. Imitation learning. Reconfigurable
systems. Supervised, non-supervised, reinforcement, and statistical algorithms.
4. Emulation of cognitive functions. Decision Making. Multi-agent systems. Sen-
sor mesh. Natural language. Pattern recognition. Perceptual and motor functions
(visual, auditory, tactile, virtual reality, etc.). Robotics. Planning motor control.
5. Bio-inspired systems and neuro-engineering. Embedded intelligent systems.
Evolvable computing. Evolving hardware. Microelectronics for neural, fuzzy, and
bioinspired systems. Neural prostheses. Retinomorphic systems. Brain–computer
interfaces (BCI). Nanosystems. Nanocognitive systems.
6. Advanced topics in computational intelligence. Intelligent networks. Knowledge-
intensive problem-solving techniques. Multi-sensor data fusion using computational
intelligence. Search and meta-heuristics. Soft Computing. Neuro-fuzzy systems.
Neuro-evolutionary systems. Neuro-swarm. Hybridization with novel computing
paradigms.
7. Applications. Expert Systems. Image and Signal Processing. Ambient intelligence.
Biomimetic applications. System identification, process control, and manufacturing.
Computational Biology and Bioinformatics. Parallel and Distributed Computing.
VI Preface

Human–Computer Interaction, Internet Modeling, Communication and Network-

ing. Intelligent Systems in Education. Human–Robot Interaction. Multi-Agent Sys-
tems. Time series analysis and prediction. Data mining and knowledge discovery.

At the end of the submission process, and after a careful peer-review and evaluation
process (each submission was reviewed by at least 2, and on the average 2.7, Program
Committee members or additional reviewers), 100 papers were accepted for oral or
poster presentation, according to the recommendations of reviewers and the authors’
preferences.
It is important to note, that for the sake of consistency and readability of the book,
the presented papers are not organized as they were presented in the IWANN 2015 ses-
sions, but classified under 14 chapters. The organization of the papers is in two volumes
arranged basically following the topics list included in the Call for Papers. The first vol-
ume (LNCS 9094), entitled “IWANN 2015. Advances on Computational Intelligence.
Part I” is divided into eight main parts and includes the contributions on:

1. Computing Languages with Bio-Inspired Devices and Multi-Agent Systems

(Special Session, organized by: M. Dolores Jiménez-López and Alfonso Ortega
de la Puente)
2. Brain-Computer Interfaces: Applications and Tele-services (Special Session,
organized by: Ricardo Ron Angevin and Miguel Angel Lopez)
3. Multi-Robot Systems: Applications and Theory (MRSAT) (Special Session,
organized by: José Guerrero and Óscar Valero)
4. Video and Image Processing (Special Session, organized by: Enrique Domínguez
and Jose Garcia)
5. Transfer Learning (Special Session, organized by: Luis M. Silva and
Jorge M. Santos)
6. Structures, algorithms, and methods in artificial intelligence (Special Session,
organized by: Daniela Danciu and Vladimir Răsvan)
7. Interactive and Cognitive Environments (Special Session, organized by: Wei Chen
and Albert Samá)
8. Mathematical and theoretical methods in Fuzzy Systems

In the second volume (LNCS 9095), entitled “IWANN 2015. Advances on Compu-
tational Intelligence. Part II” is divided into six main parts and includes the contribu-
tions on:
1. Pattern Recognition
2. Embedded intelligent systems
3. Expert Systems
4. Advances in Computational Intelligence
5. Applications of Computational Intelligence
6. Invited Talks to IWANN 2015
In this edition of IWANN 2015, the plenary talks were given by Prof.
Cristina Urdiales (The shared control paradigm for assistive and rehabilitation robots),
Prof. Dan Ciresan (Deep Neural Networks for Visual Pattern Recognition), and finally
by Prof. Andrea Cavallaro.
 Preface VII

The 13th edition of the IWANN conference was organized by the University of
Granada, University of Málaga, Polytechnic University of Catalonia, and University
of the Balearic Islands, together with the Spanish Chapter of the IEEE Computational
Intelligence Society. We wish to thank the University of the Balearic Islands for their
support and grants.
We would also like to express our gratitude to the members of the different com-
mittees for their support, collaboration, and good work. We specially thank the Local
Committe, Program Committee, the Reviewers, Invited Speaker, and Special Session
Organizers. Finally, we want to thank Springer, and especially Alfred Hoffman and
Anna Kramer for their continuous support and cooperation.

June 2015 Ignacio Rojas

Gonzalo Joya
Andreu Catala
 Organization

Program Committee
Leopoldo Acosta University of La Laguna, Spain
Vanessa Aguiar-Pulido RNASA-IMEDIR, University of A Coruña, Spain
Arnulfo Alanis Garza Instituto Tecnológico de Tijuana, Mexico
Ali Fuat Alkaya Marmara University, Turkey
Amparo Alonso-Betanzos University of A Coruña, Spain
Juan Antonio Alvarez-García University of Seville, Spain
Jhon Edgar Amaya University of Tachira (UNET), Venezuela
Gabriela Andrejkova Pavol Jozef Šafarik University, Slovak Republic
Cesar Andres Universidad Complutense de Madrid, Spain
Miguel Ángel López University of Cádiz, Spain
Anastassia Angelopoulou University of Westminster, UK
Plamen Angelov Lancaster University, UK
Davide Anguita University of Genova, Italy
Cecilio Angulo Universitat Politècnica de Catalunya, Spain
Javier Antich Universitat de les Illes Balears, Spain
Angelo Arleo CNRS - University Pierre and Marie Curie Paris VI,
France
Corneliu Arsene SC IPA SA, Romania
Miguel Atencia University of Málaga, Spain
Jorge Azorín-López University of Alicante, Spain
Davide Bacciu University of Pisa, Italy
Javier Bajo Universidad Politécnica de Madrid, Spain
Juan Pedro Bandera Rubio ISIS Group, University of Málaga, Spain
Cristian Barrué Universitat Politècnica de Catalunya, Spain
Andrzej Bartoszewicz Technical University of Lodz, Poland
Bruno Baruque University of Burgos, Spain
David Becerra Alonso University of the West of Scotland, UK
Lluís Belanche Universitat Politècnica de Catalunya, Spain
Sergio Bermejo Universitat Politècnica de Catalunya, Spain
Francesc Bonin Universitat de les Illes Balears, Spain
Francisco Bonnín Pascual Universitat de les Illes Balears, Spain
Julio Brito University of La Laguna, Spain
Antoni Burguera Universitat de les Illes Balears, Spain
Joan Cabestany Universitat Politècnica de Catalunya, Spain
Inma P. Cabrera University of Málaga, Spain
Tomasa Calvo University of Alcalá, Spain
X Organization

Jose Luis Calvo Rolle University of A Coruña, Spain

Francesco Camastra University of Naples Parthenope, Italy
Carlos Carrascosa GTI-IA DSIC Universidad Politecnica de Valencia,
Spain
Luis Castedo Universidad de A Coruña, Spain
Pedro Castillo University of Granada, Spain
Andreu Catala Universitat Politècnica de Catalunya, Spain
Ana Cavalli GET/INT, France
Miguel Cazorla University of Alicante, Spain
Wei Chen Eindhoven University of Technology,
The Netherlands
Jesus Cid-Sueiro Universidad Carlos III de Madrid, Spain
Maximo Cobos Universidad de Valencia, Spain
Valentina Colla Scuola Superiore Sant’Anna, Italy
Pablo Cordero Universidad de Málaga, Spain
Francesco Corona Aalto University, Finland
Ulises Cortes Universitat Politècnica de Catalunya, Spain
Marie Cottrell SAMM Université Paris 1 Panthéon-Sorbonne,
France
Raúl Cruz-Barbosa Universidad Tecnológica de la Mixteca, Mexico
Manuel Cruz-Ramírez University of Córdoba, Spain
Erzsébet Csuhaj-Varjú Eötvös Loránd University, Hungary
Daniela Danciu University of Craiova, Romania
Suash Deb C.V. Raman College of Engineering, India
Angel Pascual Del Pobil UNIVERSITAT JAUME I, Spain
Enrique Domínguez University of Málaga, Spain
Julian Dorado Universidade da Coruña, Spain
Abrahan Duarte Universidad Rey Juan Carlos, Spain
Richard Duro Universidad de A Coruña, Spain
Gregorio Díaz University of Castilla - La Mancha, Spain
Emil Eirola Aalto University, Finland
Patrik Eklund Umeå University, Sweden
Javier Fernández de Cañete University of Málaga, Spain
Francisco Fernandez De Vega Grupo de Evolucion Artificial,
Universidad de Extremadura, Spain
Alberto Fernandez Gil CETINIA, Universidad Rey Juan Carlos, Spain
Enrique Fernandez-Blanco University of A Coruña, Spain
Manuel Fernández Carmona Universidad de Málaga, Spain
Antonio J. Fernández Leiva Universidad de Málaga, Spain
Francisco Fernández Navarro University of Córdoba, Spain
 Organization XI

Carlos Fernández-Lozano Universidade da Coruña, Spain

Jose Manuel Ferrandez Universidad Politécnica de Cartagena, Spain
Ricardo Ferreira Nove de Julho University, Brazil
Aníbal R. Figueiras-Vidal Universidad Carlos III de Madrid, Spain
Oscar Fontenla-Romero University of A Coruña, Spain
Colin Fyfe University of the West of Scotland, UK
Emilio Garcia Universitat de les Illes Balears, Spain
Rodolfo García-Bermudez Universidad Laica Eloy Alfaro de Manabí, Ecuador
Carlos Garcia Puntonet University of Granada, Spain
Juan M Garcia-Gomez Universidad Politecnica de Valencia, Spain
Francisco Garcia-Lagos Universidad de Málaga, Spain
Jose Garcia-Rodriguez University of Alicante, Spain
Patricio García Báez Universidad de La Laguna, Spain
Pablo García Sánchez University of Granada, Spain
Maribel García-Arenas Universidad de Granada, Spain
Patrick Garda Université Pierre et Marie Curie - Paris 6, France
Peter Gloesekoetter Münster University of Applied Sciences, Germany
Juan Gomez Romero Universidad Carlos III de Madrid, Spain
Juan Gorriz University of Granada, Spain
Karl Goser Technical University Dortmund, Germany
Bernard Gosselin University of Mons, Belgium
Manuel Grana University of the Basque Country, Spain
Bertha Guijarro-Berdiñas University of A Coruña, Spain
Nicolás Guil Mata University of Málaga, Spain
Alberto Guillén University of Granada, Spain
Francisco Herrera University of Granada, Spain
Álvaro Herrero University of Burgos, Spain
Cesar Hervas University of Cordoba, Spain
Tom Heskes Radboud University Nijmegen, The Netherlands
Wei-Chiang Hong Oriental Institute of Technology, India
José M. Jerez Universidad de Málaga, Spain
M. Dolores Jimenez-Lopez Rovira i Virgili University, Spain
Juan Luis Jiménez Laredo University of Granada, Spain
Gonzalo Joya University of Málaga, Spain
Vicente Julian Universidad Politécnica de Valencia, Spain
Fernando L. Pelayo University of Castilla - La Mancha, Spain
Alberto Labarga University of Granada, Spain
Raul Lara Cabrera University of Málaga, Spain
XII Organization
Nuno Lau
Amaury Lendasse
Miguel Lopez
Otoniel López Granado
Rafael Marcos Luque Baena
Ezequiel López-Rubio
Kurosh Madani
Mario Martin
Bonifacio Martin Del Brio
Jose D. Martin-Guerrero
Luis Martí
Francisco Martínez Estudillo
José Luis Martínez Martínez
José Fco. Martínez-Trinidad
Miquel Massot
Francesco Masulli
Montserrat Mateos
Jesús Medina-Moreno
Maria Belen Melian Batista
Mercedes Merayo
Gustavo Meschino
Margaret Miro
Jose M. Molina
Augusto Montisci
Antonio Mora
Angel Mora Bonilla
Claudio Moraga
Ginés Moreno
Jose Andres Moreno
Juan Moreno Garcia
J. Marcos Moreno Vega
Susana Muñoz Hernández
Pep Lluís Negre Carrasco
Alberto Núñez
Manuel Ojeda-Aciego
Sorin Olaru
Iván Olier
Madalina Olteanu
Julio Ortega
Alfonso Ortega de La Puente
Alberto Ortiz
Emilio Ortiz-García
Universidade de Aveiro, Portugal
The University of Iowa, USA
University of Granada, Spain
Miguel Hernandez University, Spain
University of Málaga, Spain
University of Málaga, Spain
LISSI/Université PARIS-EST Creteil, France
Universitat Politècnica de Catalunya, Spain
University of Zaragoza, Spain
University of Valencia, Spain
Universidad Carlos III de Madrid, Spain
ETEA, Spain
University of Castilla-La Mancha, Spain
INAOE, Instituto Nacional de Astrofísica,
Óptica y Electrónica, Mexico
University of the Balearic Islands, Spain
University of Genoa, Italy
Universidad Pontificia de Salamanca, Spain
University of Cádiz, Morocco
University of La Laguna, Spain
Universidad Complutense de Madrid, Spain
Universidad Nacional de Mar del Plata, Spain
University of the Balearic Islands, Spain
Universidad Carlos III de Madrid, Spain
University of Cagliari, Italy
University of Granada, Spain
University of Málaga, Spain
European Centre for Soft Computing, Spain
University of Castilla-La Mancha, Spain
University of La Laguna, Spain
Universidad de Castilla-La Mancha, Spain
University of La Laguna, Spain
Technical University of Madrid, Spain
University of the Balearic Islands, Spain
Universidad de Castilla La Mancha, Spain
University of Málaga, Spain
“SUPELEC” École Supérieur d’Électricité, France
The University of Manchester, UK
SAMM, Université Paris 1, France
Universidad de Granada, Spain
Universidad Autónoma de Madrid, Spain
University of the Balearic Islands, Spain
Universidad de Alcalá, Spain
 Organization XIII

Osvaldo Pacheco Universidade de Aveiro, Portugal

Esteban José Palomo University of Málaga, Spain
Diego Pardo Barcelona Tech, Spain
Miguel Angel Patricio Universidad Carlos III de Madrid, Spain
Alejandro Pazos Sierra University of A Coruña, Spain
Jose Manuel Perez Lorenzo Universidad de Jaén, Spain
Vincenzo Piuri University of Milan, Italy
Héctor Pomares University of Granada, Spain
Alberto Prieto Universidad de Granada, Spain
Alexandra Psarrou University of Westminster, UK
Francisco A. Pujol University of Alicante, Spain
Pablo Rabanal Universidad Complutense de Madrid, Spain
Juan Rabuñal University of A Coruña, Spain
Vladimir Răsvan Universitatea din Craiova, Romania
Ismael Rodríguez Universidad Complutense de Madrid, Spain
Juan A. Rodriguez Universidad de Málaga, Spain
Sara Rodríguez University of Salamanca, Spain
Fernando Rojas University of Granada, Spain
Ignacio Rojas University of Granada, Spain
Samuel Romero-Garcia University of Granada, Spain
Ricardo Ron-Angevin University of Málaga, Spain
Eduardo Ros University of Granada, Spain
Francesc Rossello University of the Balearic Islands, Spain
Fabrice Rossi SAMM, Université Paris 1, France
Fernando Rubio Universidad Complutense de Madrid, Spain
Ulrich Rueckert University of Paderborn, Germany
Addisson Salazar Universidad Politécnica de Valencia, Spain
Sancho Salcedo-Sanz Universidad de Alcalá, Spain
Albert Samà Universitat Politècnica de Catalunya, Spain
Francisco Sandoval Universidad de Málaga, Spain
José Santos University of A Coruña, Spain
Jose A. Seoane University of Bristol, UK
Eduardo Serrano Universidad Autónoma de Madrid, Spain
Luís Silva University of Aveiro, Portugal
Olli Simula Helsinki University of Technology, Finland
Jordi Solé-Casals Universitat de Vic, Spain
Carmen Paz Suárez Araujo Universidad de las Palmas de Gran Canaria, Spain
Peter Szolgay Pázmány Péter Catholic University, Hungary
Javier Sánchez-Monedero University of Córdoba, Spain
Ricardo Tellez Pal Robotics, Spain
Ana Maria Tome Universidade de Aveiro, Portugal
Carme Torras IRI (CSIC-UPC), Spain
Joan Torrens University of the Balearic Islands, Spain
XIV Organization

Claude Touzet University of Provence, France

Olga Valenzuela University of Granada, Spain
Óscar Valero University of the Balearic Islands, Spain
Miguel Ángel Veganzones Universidad del País Vasco (UPV/EHU), Spain
Francisco Velasco-Álvarez Universidad de Málaga, Spain
Sergio Velastin Kingston University, UK
Marley Vellasco PUC-Rio, Brazil
Alfredo Vellido Universitat Politècnica de Catalunya, Spain
Francisco J. Veredas Universidad de Málaga, Spain
Michel Verleysen Université catholique de Louvain, Belgium
Changjiu Zhou Singapore Polytechnic, Singapore
Ahmed Zobaa University of Exeter, UK

Additional Reviewers
Azorín-López, Jorge Navarro-Ortiz, Jorge
Cortes, Ulises Ortiz, Alberto
De La Cruz, Marina Palomo, Esteban José
Fernandez-Blanco, Enrique Peters, Peter
Garcia-Fidalgo, Emilio Rodriguez, Juan A.
Georgieva, Petia Rodriguez-Benitez, Luis
Luque-Baena, Rafael M. Sánchez-Morillo, Daniel
López-Rubio, Ezequiel Volosyak, Ivan
Martinez-Gomez, Jesus Wang, Qi
Moreno Garcia, Juan Wetzels, Mart
 Contents – Part II

Pattern Recognition

Developing Gene Classifier System for Autism Recognition . . . . . . . . . . . . 3

Tomasz Latkowski and Stanislaw Osowski

A Distributed Feature Selection Approach Based on a Complexity Measure. . . 15

Verónica Bolón-Canedo, Noelia Sánchez-Maroño,
and Amparo Alonso-Betanzos

Ensemble Feature Selection for Rankings of Features . . . . . . . . . . . . . . . . . 29

Borja Seijo-Pardo, Verónica Bolón-Canedo, Iago Porto-Díaz,
and Amparo Alonso-Betanzos

A Medical Case-Based Reasoning Approach Using Image Classification

and Text Information for Recommendation . . . . . . . . . . . . . . . . . . . . . . . . 43
Sara Nasiri, Johannes Zenkert, and Madjid Fathi

Non Spontaneous Saccadic Movements Identification in Clinical

Electrooculography Using Machine Learning . . . . . . . . . . . . . . . . . . . . . . . 56
Roberto Becerra-García, Rodolfo García-Bermúdez,
Gonzalo Joya-Caparrós, Abel Fernández-Higuera,
Camilo Velázquez-Rodríguez, Michel Velázquez-Mariño,
Franger Cuevas-Beltrán, Francisco García-Lagos,
and Roberto Rodríguez-Labrada

Applying a Hybrid Algorithm to the Segmentation of the Spanish Stock

Market Index Time Series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 69
Antonio Manuel Durán-Rosal, Mónica de la Paz-Marín,
Pedro Antonio Gutiérrez, and César Hervás-Martínez

Nonlinear Ordinal Logistic Regression Using Covariates Obtained

by Radial Basis Function Neural Networks Models. . . . . . . . . . . . . . . . . . . 80
Manuel Dorado-Moreno, Pedro Antonio Gutiérrez,
Javier Sánchez-Monedero, and César Hervás-Martínez

Energy Flux Range Classification by Using a Dynamic Window

Autoregressive Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92
Pedro Antonio Gutiérrez, Juan Carlos Fernández, Mária Pérez-Ortiz,
Laura Cornejo-Bueno, Enrique Alexandre-Cortizo, Sancho Salcedo-Sanz,
and César Hervás-Martínez
XVI Contents – Part II

Automatic Eye Blink Detection Using Consumer Web Cameras. . . . . . . . . . 103

Beatriz Remeseiro, Alba Fernández, and Madalena Lira

Insights on the Use of Convolutional Neural Networks for Document

Image Binarization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 115
J. Pastor-Pellicer, S. España-Boquera, F. Zamora-Martínez,
M. Zeshan Afzal, and Maria Jose Castro-Bleda

A Genetic Algorithms-Based LSSVM Classifier for Fixed-Size

Set of Support Vectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
Danilo Avilar Silva and Ajalmar R. Rocha Neto

Ensemble of Minimal Learning Machines for Pattern Classification . . . . . . . 142

Diego Parente Paiva Mesquita, João Paulo Pordeus Gomes,
and Amauri Holanda Souza Junior

Extreme Learning Machines for Multiclass Classification: Refining

Predictions with Gaussian Mixture Models . . . . . . . . . . . . . . . . . . . . . . . . 153
Emil Eirola, Andrey Gritsenko, Anton Akusok, Kaj-Mikael Björk,
Yoan Miche, Dušan Sovilj, Rui Nian, Bo He, and Amaury Lendasse

Modeling the EUR/USD Index Using LS-SVM and Performing

Variable Selection. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 165
Luis-Javier Herrera, Alberto Guillén, Rubén Martínez, Carlos García,
Hector Pomares, Oresti Baños, and Ignacio Rojas

Embedded Intelligent Systems

Modeling Retina Adaptation with Multiobjective Parameter Fitting . . . . . . . 175

Pablo Martínez-Cañada, Christian Morillas, Samuel Romero,
and Francisco Pelayo

Stochastic-Based Implementation of Reservoir Computers . . . . . . . . . . . . . . 185

Miquel L. Alomar, Vincent Canals, Víctor Martínez-Moll,
and Josep L. Rosselló

FPGA Implementation Comparison Between C-Mantec

and Back-Propagation Neural Network Algorithms . . . . . . . . . . . . . . . . . . . 197
Francisco Ortega-Zamorano, José M. Jerez, Gustavo Juárez,
and Leonardo Franco

Expert Systems

Logic Programming and Artificial Neural Networks in Breast Cancer

Detection . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
José Neves, Tiago Guimarães, Sabino Gomes, Henrique Vicente,
Mariana Santos, João Neves, José Machado, and Paulo Novais
 Contents – Part II XVII

An ANFIS-Based Fault Classification Approach in Double-Circuit

Transmission Line Using Current Samples. . . . . . . . . . . . . . . . . . . . . . . . . 225
Mohammad Amin Jarrahi, Haidar Samet, Hossein Raayatpisheh,
Ahmad Jafari, and Mohsen Rakhshan

Evolutionary Hybrid Configuration Applied to a Polymerization Process

Modelling . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 237
Silvia Curteanu, Elena-Niculina Dragoi, and Vlad Dafinescu

Multi-layer Perceptrons for Voxel-Based Classification of Point Clouds

from Natural Environments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 250
Victoria Plaza, Jose Antonio Gomez-Ruiz, Anthony Mandow,
and Alfonso J. Garcia-Cerezo

An Improved RBF Neural Network Approach to Nonlinear Curve Fitting . . . 262

Michael M. Li and Brijesh Verma

QSVM: A Support Vector Machine for Rule Extraction . . . . . . . . . . . . . . . 276

Guido Bologna and Yoichi Hayashi

Multiwindow Fusion for Wearable Activity Recognition . . . . . . . . . . . . . . . 290

Oresti Baños, Juan-Manuel Galvez, Miguel Damas, Alberto Guillén,
Luis-Javier Herrera, Hector Pomares, Ignacio Rojas, Claudia Villalonga,
Choong Seon Hong, and Sungyoung Lee

Ontological Sensor Selection for Wearable Activity Recognition . . . . . . . . . 298

Claudia Villalonga, Oresti Baños, Hector Pomares, and Ignacio Rojas

Short-Term Spanish Aggregated Solar Energy Forecast . . . . . . . . . . . . . . . . 307

Nicolas Perez-Mora, Vincent Canals, and Víctor Martínez-Moll

Intelligent Presentation Skills Trainer Analyses Body Movement . . . . . . . . . 320

Anh-Tuan Nguyen, Wei Chen, and Matthias Rauterberg

Performing Variable Selection by Multiobjective Criterion: An Application

to Mobile Payment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 333
Alberto Guillén, Luis-Javier Herrera, Francisco Liébana, Oresti Baños,
and Ignacio Rojas

Advances in Computational Intelligence

Aggregation of Partial Rankings – An Approach Based on the Kemeny

Ranking Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 343
Gonzalo Nápoles, Zoumpoulia Dikopoulou, Elpiniki Papageorgiou,
Rafael Bello, and Koen Vanhoof

Existence and Synthesis of Complex Hopfield Type Associative Memories . . . 356

Garimella Rama Murthy and Moncef Gabbouj
XVIII Contents – Part II

On Acceleration of Incremental Learning in Chaotic Neural Network . . . . . . 370

Toshinori Deguchi, Toshiki Takahashi, and Naohiro Ishii

Comparing Optimization Methods, in Continuous Space, for Modelling

with a Diffusion Process . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 380
Nuria Rico, Maribel García Arenas, Desirée Romero,
J.M. Crespo, Pedro Castillo, and J.J. Merelo

Estimating Artificial Neural Networks with Generalized Method Moments . . . 391

Alexandre Street de Aguiar and João Marco Braga da Cunha

An Hybrid Ensemble Method Based on Data Clustering

and Weak Learners Reliabilities Estimated Through Neural Networks . . . . . 400
Marco Vannucci, Valentina Colla, and Silvia Cateni

Conventional Prediction vs Beyond Data Range Prediction

of Loss Coefficient for Quarter Circle Breakwater Using ANFIS . . . . . . . . . 412
Arkal Vittal Hegde and Raju Budime

Performance Evaluation of Least Squares SVR in Robust Dynamical

System Identification. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 422
José Daniel A. Santos, César Lincoln C. Mattos,
and Guilherme A. Barreto

On the Generalization of the Uninorm Morphological Gradient . . . . . . . . . . 436

Manuel González-Hidalgo, Sebastia Massanet, Arnau Mir,
and Daniel Ruiz-Aguilera

Revisiting Image Vignetting Correction by Constrained Minimization

of Log-Intensity Entropy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 450
Laura Lopez-Fuentes, Gabriel Oliver, and Sebastia Massanet

Hybrid Dynamic Learning Systems for Regression . . . . . . . . . . . . . . . . . . . 464

Kaushala Dias and Terry Windeatt

A Novel Algorithm to Train Multilayer Hardlimit Neural Networks Based

on a Mixed Integer Linear Program Model . . . . . . . . . . . . . . . . . . . . . . . . 477
Jose B. da Fonseca

On Member Labelling in Social Networks . . . . . . . . . . . . . . . . . . . . . . . . . 488

Rafael Corchuelo, Antonia M. Reina Quintero, and Patricia Jiménez

Applications of Computational Intelligence

Deconvolution of X-ray Diffraction Profiles Using Genetic Algorithms

and Differential Evolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 503
Sidolina P. Santos, Juan A. Gomez-Pulido, and Florentino Sanchez-Bajo
 Contents – Part II XIX

Using ANN in Financial Markets Micro-Structure Analysis . . . . . . . . . . . . . 515

Brayan S. Reyes Daza and Octavio J. Salcedo Parra

Cluster Analysis of Finger-to-nose Test for Spinocerebellar Ataxia

Assessment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 524
Michel Velázquez-Mariño, Miguel Atencia, Rodolfo García-Bermúdez,
Daniel Pupo-Ricardo, Roberto Becerra-García, Luis Velázquez Pérez,
and Francisco Sandoval

Exploiting Neuro-Fuzzy System for Mobility Prediction in Wireless

Ad-Hoc Networks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 536
Mohamed Elleuch, Heni Kaaniche, and Mohamed Ayadi

A New Method for an Optimal SOM Size Determination in Neuro-Fuzzy

for the Digital Forensics Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . 549
Andrii Shalaginov and Katrin Franke

SVRs and Uncertainty Estimates in Wind Energy Prediction . . . . . . . . . . . . 564

Jesús Prada and José Ramón Dorronsoro

Search for Meaning Through the Study of Co-occurrences in Texts . . . . . . . 578

Nicolas Bourgeois, Marie Cottrell, Stéphane Lamassé,
and Madalina Olteanu

Evaluation of Fitting Functions for the Saccade Velocity Profile

in Electrooculographic Records . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 592
Rodolfo García-Bermúdez, Camilo Velázquez-Rodríguez, Fernando Rojas,
Manuel Rodríguez, Roberto Becerra-García, Michel Velázquez-Mariño,
José Arteaga-Vera, and Luis Velázquez

esCam: A Mobile Application to Capture and Enhance Text Images. . . . . . . 601

J. Pastor-Pellicer, Maria Jose Castro-Bleda, and J.L. Adelantado-Torres

Computer Access and Alternative and Augmentative Communication

(AAC) for People with Disabilities: A Multi-modal Hardware
and Software Solution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 605
Salvador Sancha-Ros and Esther García-Garaluz

Invited Talks to IWANN 2015

The Shared Control Paradigm for Assistive and Rehabilitation Robots . . . . . 613
Cristina Urdiales

Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 617

 Contents – Part I

Computing Languages with Bio-Inspired Devices

and Multi-Agent Systems

A Grammatical Inference Model for Measuring Language Complexity . . . . . 3

Leonor Becerra-Bonache and M. Dolores Jiménez-López

A Proposal for Contextual Grammatical Inference . . . . . . . . . . . . . . . . . . . 18

Leonor Becerra-Bonache, María Galván, and François Jacquenet

How to Search Optimal Solutions in Big Spaces with Networks

of Bio-Inspired Processors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29
José Ramón Sánchez Couso, Sandra Gómez Canaval,
and David Batard Lorenzo

Distributed Simulation of NEPs Based On-Demand Cloud Elastic

Computation. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
Sandra Gómez Canaval, Alfonso Ortega de la Puente,
and Pablo Orgaz González

How Nets of Evolutionary Processors (NEPs) Could be Simulated

in a Distributed Way . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
Karina Jiménez, Antonio Jiménez, Marina de la Cruz,
and Sandra Gómez Canaval

Brain-Computer Interfaces: Applications and Tele-services

A Comparison of SSVEP-Based BCI-Performance Between Different

Age Groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
Felix Gembler, Piotr Stawicki, and Ivan Volosyak

Training in Realistic Virtual Environments: Impact on User Performance

in a Motor Imagery-Based Brain–Computer Interface . . . . . . . . . . . . . . . . . 78
Leandro da Silva-Sauer, Luis Valero-Aguayo, Francisco Velasco-Álvarez,
Sergio Varona-Moya, and Ricardo Ron-Angevin

Real-Time Monitoring of Biomedical Signals to Improve Road Safety . . . . . 89

José Miguel Morales, Leandro Luigi Di Stasi, Carolina Díaz-Piedra,
Christian Morillas, and Samuel Romero
XXII Contents – Part I

Brain-Computer Interface: Usability Evaluation of Different P300 Speller

Configurations: A Preliminary Study. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98
Liliana Garcia, Véronique Lespinet-Najib, Sarah Saioud,
Victor Meistermann, Samuel Renaud, Jaime Diaz-Pineda,
Jean Marc André, and Ricardo Ron-Angevin

Accessing Tele-Services Using a Hybrid BCI Approach . . . . . . . . . . . . . . . 110

Chris Brennan, Paul McCullagh, Gaye Lightbody, Leo Galway,
Diana Feuser, José Luis González, and Suzanne Martin

Authentication of Brain-Computer Interface Users in Network Applications. . . 124

M.A. Lopez-Gordo, Ricardo Ron-Angevin, and Francisco Pelayo

A Label-Aided Filter Method for Multi-objective Feature Selection

in EEG Classification for BCI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
Pedro Martín-Smith, Julio Ortega, Javier Asensio-Cubero,
John Q. Gan, and Andrés Ortiz

Multi-Robot Systems: Applications and Theory (MRSAT)

A First Step Toward a Possibilistic Swarm Multi-robot Task Allocation . . . . 147

José Guerrero, Óscar Valero, and Gabriel Oliver

A Bottom-up Robot Architecture Based on Learnt Behaviors Driven Design . . . 159

Ignacio Herrero, Cristina Urdiales García, José Manuel Peula Palacios,
and Francisco Sandoval Hernández

From Human Eye Fixation to Human-like Autonomous Artificial Vision . . . 171

Viachaslau Kachurka, Kurosh Madani, Cristophe Sabourin,
and Vladimir Golovko

Towards a Shared Control Navigation Function: Efficiency Based

Command Modulation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 185
Manuel Fernández-Carmona, José Manuel Peula,
Cristina Urdiales, and Francisco Sandoval

AMiRo: A Mini Robot for Scientific Applications . . . . . . . . . . . . . . . . . . . 199

Thomas Schöpping, Timo Korthals, Stefan Herbrechtsmeier,
and Ulrich Rückert

Video and Image Processing

Visualization of Complex Datasets with the Self-Organizing Spanning Tree. . . 209

Ezequiel López-Rubio, Esteban José Palomo,
Rafael Marcos Luque Baena, and Enrique Domínguez
 Contents – Part I XXIII

A Detection System for Vertical Slot Fishways Using Laser Technology

and Computer Vision Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218
Angel J. Rico-Diaz, Alvaro Rodriguez, Daniel Villares,
Juan R. Rabuñal, Jeronimo Puertas, and Luis Pena

Interactive Relevance Visual Learning for Image Retrieval . . . . . . . . . . . . . 227

Hsin-Chia Fu, Z.H. Wang, W.J. Wang, and Hsiao-Tien Pao

Scene Classification Based on Local Binary Pattern and Improved Bag

of Visual Words . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 241
Gholam Ali Montazer, Davar Giveki, and Mohammad Ali Soltanshahi

An Experimental Comparison for the Identification of Weeds in Sunflower

Crops via Unmanned Aerial Vehicles and Object-Based Analysis. . . . . . . . . 252
María Pérez-Ortiz, Pedro Antonio Gutiérrez, Jose Manuel Peña,
Jorge Torres-Sánchez, César Hervás-Martínez,
and Francisca López-Granados

A Novel Framework for Hyperemia Grading Based on Artificial

Neural Networks. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263
Luisa Sánchez, Noelia Barreira, Hugo Pena-Verdeal,
and Eva Yebra-Pimentel

Applying a Genetic Algorithm Solution to Improve Compression

of Wavelet Coefficient Sign . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 276
Antonio Martí, Otoniel López, Francisco Rodríguez-Ballester,
and Manuel Malumbres

Finding the Texture Features Characterizing the Most Homogeneous

Texture Segment in the Image . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 287
Alexander Goltsev, Vladimir Gritsenko, Ernst Kussul, and Tatiana Baidyk

Robust Tracking for Augmented Reality . . . . . . . . . . . . . . . . . . . . . . . . . . 301

José M. González-Linares, Nicolás Guil, and Julián Ramos Cózar

Bio-inspired Motion Estimation with Event-Driven Sensors . . . . . . . . . . . . . 309

Francisco Barranco, Cornelia Fermuller, and Yiannis Aloimonos

Transfer Learning

Domain Generalization Based on Transfer Component Analysis. . . . . . . . . . 325

Thomas Grubinger, Adriana Birlutiu, Holger Schöner,
Thomas Natschläger, and Tom Heskes

Deep Transfer Learning Ensemble for Classification . . . . . . . . . . . . . . . . . . 335

Chetak Kandaswamy, Luís M. Silva, Luís A. Alexandre,
and Jorge M. Santos
XXIV Contents – Part I

Development of a Power Output Forecasting Tool for Wind Farms

Based in Principal Components and Artificial Neural Networks . . . . . . . . . . 349
P. del Saz-Orozco, J. Fernández de Cañete, and R. Alba

CO2 RBFN-CS: First Approach Introducing Cost-Sensitivity

in the Cooperative-Competitive RBFN Design . . . . . . . . . . . . . . . . . . . . . . 361
María Dolores Pérez-Godoy, Antonio Jesús Rivera, Francisco Charte,
and Maria Jose del Jesus

Transfer Learning for the Recognition of Immunogold Particles

in TEM Imaging. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 374
Ricardo Gamelas Sousa, Tiago Esteves, Sara Rocha,
Francisco Figueiredo, Joaquim Marques de Sá, Luís A. Alexandre,
Jorge M. Santos, and Luís M. Silva

Structures, Algorithms and Methods in Artificial Intelligence

BSO-FS: Bee Swarm Optimization for Feature Selection in Classification. . . 387

Souhila Sadeg, Leila Hamdad, Karima Benatchba, and Zineb Habbas

Improved Retrieval for Challenging Scenarios in Clique-Based

Neural Networks. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 400
Xiaoran Jiang, Max Raphael Sobroza Marques, Pierre-Julien Kirsch,
and Claude Berrou

On Structures with Emergent Computing Properties. A Connectionist versus

Control Engineering Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 415
Daniela Danciu and Vladimir Răsvan

Deep Neural Networks for Wind Energy Prediction . . . . . . . . . . . . . . . . . . 430

David Díaz, Alberto Torres, and José Ramón Dorronsoro

Ensemble of Classifiers for Length of Stay Prediction in Colorectal Cancer . . . 444

Ruxandra Stoean, Catalin Stoean, Adrian Sandita, Daniela Ciobanu,
and Cristian Mesina

Interactive and Cognitive Environments

Monitoring Motor Fluctuations in Parkinson’s Disease Using a Waist-Worn

Inertial Sensor . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 461
Carlos Pérez-López, Albert Samà, Daniel Rodríguez-Martín,
Andreu Català, Joan Cabestany, Eva de Mingo,
and Alejandro Rodríguez-Molinero
 Contents – Part I XXV

Convolutional Neural Networks for Detecting and Mapping Crowds

in First Person Vision Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 475
Juan Sebastian Olier, Carlo Regazzoni, Lucio Marcenaro,
and Matthias Rauterberg

E-COmate: What’s Your Non-consumption? . . . . . . . . . . . . . . . . . . . . . . . 486

Veranika Lim, Mathias Funk, Matthias Rauterberg, Lucio Marcenaro,
and Carlo Regazzoni

Mathematical and Theoretical Methods in Fuzzy Systems

Extended Bag of Visual Words for Face Detection . . . . . . . . . . . . . . . . . . . 503

Gholam Ali Montazer, Mohammad Ali Soltanshahi, and Davar Giveki

Improving Multi-adjoint Logic Programs by Unfolding Fuzzy Connective

Definitions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 511
Pedro J. Morcillo and Ginès Moreno

A Mixed Fuzzy Similarity Approach to Detect Plagiarism in Persian Texts . . . 525

Hamid Ahangarbahan and Gholam Ali Montazer

A Neural-Network-Based Robust Observer for Simultaneous Unknown

Input Decoupling and Fault Estimation . . . . . . . . . . . . . . . . . . . . . . . . . . . 535
Piotr Witczak, Marcin Mrugalski, Krzysztof Patan, and Marcin Witczak

Consequences of Structural Differences Between Hierarchical Systems

While Fuzzy Inference . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 549
Begum Mutlu, Ebru A. Sezer, and Hakan A. Nefeslioglu

SIRMs Fuzzy Inference Model with Linear Transformation of Input

Variables and Universal Approximation. . . . . . . . . . . . . . . . . . . . . . . . . . . 561
Hirofumi Miyajima, Noritaka Shigei, and Hiromi Miyajima

A New Approach of Fuzzy Neural Networks in Monthly Forecast

of Water Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 576
Ruben Araújo, Mêuser Valença, and Sérgio Fernandes

Ordering Relations over Intuitionistic Fuzzy Quantities . . . . . . . . . . . . . . . . 587

Elena Mielcova

On Fuzzy c-Means and Membership Based Clustering . . . . . . . . . . . . . . . . 597

Vicenç Torra

Author Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 609

Pattern Recognition
 Developing Gene Classifier System
for Autism Recognition

Tomasz Latkowski1 and Stanislaw Osowski1,2 ()

1
Warsaw University of Technology, Warsaw, Poland
tlatkowski@wat.edu.pl, sto@iem.pw.edu.pl
2
Military University of Technology, Warsaw, Poland

Abstract. The paper presents comparison of few chosen approaches to recogni-

tion of autism on the basis of gene expression microarray. The important point
in this task is selection of genes of the highest class discriminative ability. To
solve the problem we have applied many selection methods, which are based on
different principles. The limited set of genes in each method are selected for
further analysis. In this paper we will compare the genetic algorithm and ran-
dom forest in the role of final gene selection. The most important genes selected
by each method are used as the input attributes to the support vector machine
and random forest classifiers, cooperating in an ensemble. The final result of
classification is generated by the random forest, performing the role of fusion
system for an ensemble.

Keywords: Autism · Ensemble of classifiers · Gene selection · SVM · Random

forest

1 Introduction
Autism spectrum disorder belongs to severe neurodevelopmental disorders with social
and communication development [1]. The idea of using gene expression in autism
recognition is relatively new and underexplored now [2]. The research is directed to
identifying limited group of genes, which are strongly related to disease and allow
recognizing the autistic data from the reference samples. The main problem is an ill-
conditioning of the gene expression matrix. The number of genes is in the range of
few dozens of thousands and number of observations (patients) very scarce (usually in
the range of one hundred). Application of classical feature selection methods leads to
poor results, not adequate to the real association of the genes with the class member-
ship of data.
Actual approaches to the gene selection are concerned mainly with different types
of cancer and apply many different methods, including neural networks and Support
Vector Machines [3,4], linear regression methods [5], statistical tests [6,7], rough set
theory [8], genetic algorithms [9], frequent itemset mining [10], as well as a combina-
tion of many selection methods [11,12].
In this paper we propose the application of multistage selection procedure to aut-
ism recognition. In the first stage we reduce the population of genes by eliminating

© Springer International Publishing Switzerland 2015

I. Rojas et al. (Eds.): IWANN 2015, Part II, LNCS 9095, pp. 3–14, 2015.
DOI: 10.1007/978-3-319-19222-2_1
4 T. Latkowski and S. Osowski

the genes, which have very similar expression values for the autistic and reference
(healthy) classes. In this way it is possible to halve the size of genes. In the next stage
the genes are subject to selection according to the assumed measure of quality. Eight
different methods of selection are applied in this stage. They include both wrapper
and filter methods, such as Fisher discriminant, correlation of feature with a class,
three statistical hypothesis tests, reliefF algorithm, application of recursive linear
SVM feature elimination, and stepwise regression method [3]. The genes are ordered
according to their discrimination ability related to the particular method. In further
steps of analysis we limit their population to 100 the best. The important problem in
this stage is to find the small number of the globally best genes, chosen as the most
important in the repeated runs of the selecting procedure.
We will compare the application of few different approaches. In the first one we
use the genetic algorithm [13] responsible for the selection of the most significant
genes in the process of classification. The genetic algorithm determines automatically
the optimal size of genes. In the next approach we use the random forest (RF) [14]
applied sequentially for selection of the most important genes. The ordering of genes
in this method depends on how the particular genes influence the process of classifi-
cation. Their optimal number is determined by checking the accuracy of classifica-
tion, corresponding to the particular population of genes.
As a result of such procedure we select eight different sets of genes, treated as the
most significant in classification. Each of these sets represents the input attributes for
the support vector machine (SVM) [15] and RF classifiers, forming an ensemble. The
final decision of recognizing the autistic case from the reference one is done by fusing
the outputs of the ensemble members. This is done by the random forest (RF), serving
this time the role of integrator. For comparison we have checked also the direct appli-
cation of best sets of genes selected by all methods as the input attributes to the ran-
dom forest serving simultaneously the role of classifiers and an integrator.
The numerical experiments were performed on the NCBI base of autism [16]. They
have shown very encouraging results. The classification results of application of indi-
vidual methods have been significantly improved by such approach.

2 Gene Selection Methods

2.1 Data Base of Autism

In the numerical experiments we have used a publicly available database in GEO
(NCBI) repository [16]. It contains 146 observations (patients) and 54613 genes.
Among the patients there were 82 cases related to children with autism (treated as
class 1) and 64 of healthy children forming the reference group (class 2). All subjects
were males. Total RNA was extracted with Affymetrix Human U133 Plus 2.0 39
Expression Arrays. Children with autism were diagnosed by a medical professional
according to the DSM-IV criteria and the diagnosis was confirmed by ADOS and
ADI-R criteria [17]. The study population was primarily Caucasian and there were no
group level differences in ethnicity.
 Developing Gene Classifier System for Autism Recognition 5

2.2 Initial Step

The genes, regarded as important in class recognition, should show different expres-
sion value in class 1 and class 2. In the introductory step we use the median of their
class value within all observations. Taking into account very high variance of these
expressions within the same class we must be very careful in selecting the threshold.
For example in the data base [16] the mean values of gene expression of the members
of the autistic class extended from the smallest 4.96 to the largest 19000. At the same
time standard deviations have varied up 1138. The curve presented in Fig. 1 shows
that most genes depicts very similar ratio of the medians in both classes.

Fig. 1. The curve representing the population of eliminated genes at their differing median ratio
in two classes

It is seen that there are no genes of the ratio below the value of 70%. The active
range of reduction is from 80% to 100%. In our experiments we set the threshold on
the level of 0.96. This value was chosen heuristically to reduce the number of less
significant genes to a reasonable level. It resulted in elimination of 35762 genes.
Therefore in further analysis we used only 18851 genes, for which the ratio of me-
dians in both classes was below 0.96.

2.3 Selection Methods in the First Stage

The reduced set of genes was subject to further reduction by applying eight feature
selection methods. They were chosen in a way to provide their highest possible statis-
tical independence. The following methods have been selected: Fisher discriminant,
correlation of feature with a class, three statistical hypothesis tests (Kolmogorov-
Smirnov test, two-sample Student t-test and Kruskal-Wallis test), reliefF algorithm,
application of recursive linear SVM feature elimination, and stepwise regression me-
thod [3].
In Fisher test the genes were selected according to the 2-class discrimination
measure
6 T. Latkowski and S. Osowski

c1 − c2
S12 ( g ) = (1)
σ1 + σ 2
in which where c1 and c2 are the mean values for classes 1 and 2, respectively, while
σ1 and σ2 the appropriate standard deviations. A large value of this measure indicates
good class discriminative ability of the gene.
In correlation method we examine the correlation of gene g with a class. For two
classes of data the correlation measure is given in the form
2
 Pk (ck − c) 2
k =1 (2)
S (g) = 2
σ 2  Pk (1 − Pk )
k =1

where c and σ represent the mean value and standard deviation of gene g for all data,
ck is a mean value of the gene expression for the kth class data and Pk is a probability
of kth class occurrence in the data set.
The statistical hypothesis tests applied in experiments compare the medians of the
groups of data to determine if the samples come from the same population. Kolmogo-
rov-Smirnov test is based on the distance between the cumulative distribution of sam-
ples F1(x) and F2(x) of the gene belonging to class 1 and 2. In Kruskal-Wallis test the
ranks of the data rather than the numeric values of gene expressions are used. High
distances in both tests mean good recognition ability of the particular gene. In a two-
sample Student t-test the null hypothesis is checked if the data in the class 1 and 2 are
independent random samples of normal distributions with equal means and equal, but
unknown variances, against the alternative hypothesis that the means are not equal.
Fulfilment of the alternative hypothesis means good class discrimination ability of the
gene. The details of these statistics can be found in [18].
The reliefF algorithm checks the discrimination ability of the gene according to its
highest correlation with the observed class while taking into account the distances
between opposite classes. The quality of the gene is assessed according to how well
its expression values distinguish between observations that are near to both opposite
classes.
Stepwise linear regression is a systematic method based on adding and removing
features to the set of input attributes according to their statistical significance in a
regression. It compares the explanatory power of incrementally larger and smaller
models based on the p value of F-statistics [18,19]. Based on the statistical result a
decision is taken whether the gene should be included in a model or not.
In the SVM-recursive feature elimination method the linear SVM network is
trained at application of all genes. The decision of accepting or removing the particu-
lar gene is taken on the basis of value of weight joining the gene with a SVM classifi-
er. High weight means significant influence of the gene. The features are eliminated
step by step, since the SVM is retrained at each step with the population of features
becoming gradually smaller and smaller.
Another random document with
no related content on Scribd:
have I been wounded. But I have been knocked over frequently
through carelessness in approaching boar at bay down-wind, or in
stalking at night. The latter sport, especially when stalking a solitaire,
is very exciting: it requires skill, patience, and great caution.
I wear, when stalking, shoes with rope-soles, enabling me to tread
noiselessly over rough ground. I have stalked boar on a dark night
up-wind, when feeding in corn, until I have approached the animal
hidden by the crop, and have put the barrel of my gun within a foot of
his body before firing. When I heard the boar occupied in tearing off
a pod of Indian corn or munching grain, I advanced. When he
stopped feeding to listen, as they will cunningly do for several
minutes, I stood motionless also, until the munching recommenced.
One very dark night I managed to approach so noiselessly along
a narrow path through a copse which led to an orchard—where I had
heard from the windows of my villa at ‘Ravensrock’ a boar eating
apples—that I actually pushed my knee against the boar, who had
his snout in an opposite direction, before either of us became aware
that we were at close quarters. My gun was not cocked, for I did not
expect to have to use it until I entered the orchard, where I supposed
the boar to be still feeding. The leap I made in the air was not more
frantic than that of the boar, who jumped into the thicket. We were
both terribly startled. The boar had no doubt in the still night heard
me close the door of the balcony, two hundred yards off from the
orchard, and had hidden in the dark path to listen and await events.
On another occasion, having observed during my rides on the hill
that boar came down at night to a rough field of barley, I took my gun
a little before sunset and rode to the ground. I left my nag in charge
of a Moor, about a quarter of a mile from the field, and directed him
to keep quiet, and not to come near the field until I fired a shot. The
crop of barley I had observed was poor and short, so I felt sure I
should see the body of any boar worth firing at.
I seated myself on a rock about three feet from the ground. In my
belt was a long Spanish knife, with a handle made to insert in the
muzzle of a gun, like a bayonet. The moon had set, the sky was
cloudy, and starlight very faint. I wrapped a piece of white paper as a
sight around the gun, a few inches from the lock, so that I could see
it, even though the night was very dark. Just as the nine p.m.
Gibraltar gun boomed across the Straits, I heard a rustling in the
bush and a grunt, warning me the enemy was nigh.
The wind was favourable; the boar had entered the field on a
different side from what I expected. I strained my eyes to view the
beast, whom I could hear chewing the ears of barley, but could not at
first distinguish him.
At length he approached within fifteen yards from the rock where I
was seated, and I could just see his head above the barley, therefore
I concluded, supposing the stalk was short, it was a sow or only a
two-year-old. I waited until the object advanced within a few yards,
and I could see a good patch of black body. I fired, and heard the
noise of the fall; then the boar rose, went a few yards, and tumbled
over, and I could distinctly hear what appeared to be its death-
struggles. Then all was still; I got down from the rock, but did not
reload, thinking there was no risk, and walked to the spot where I
heard the struggles.
In the short barley were several low palmetto bushes. Seeing a
dark object move, as I fancied, I aimed and fired. It was a palmetto
bush—the leaves shaken by the wind had rustled. Within a few yards
of this bush a large form suddenly rose and came slowly towards
me. Both barrels were empty. I had barely time to insert the Spanish
knife in the muzzle of the gun when I could see a grim head and
tusks glistening in the starlight. It was not, as I had supposed, a sow
or a pig; it was a tusker.
The ground was favourable, for I stood uphill above the boar. I
held the gun so that the knife should enter at the shoulder and not
strike the head. As the boar pressed on to reach me, I joyfully felt the
blade penetrate into its body up to the hilt, and expected he would
fall dead; but no, limping on one sound leg he continued to advance;
so I backed, nearly falling over a palmetto bush; then the boar
moved to one side to get round upon me, and I followed his
movements, dreading every moment that the knife, if the boar
retreated, would be withdrawn.
Again he came on with a rush, and I moved rapidly backwards
until my back came against a rock in the field about four feet from
the ground. I scrambled up it, pressing the knife and gun against the
boar’s body to assist me. He tried to follow, but, with his disabled leg,
failed and then moved away, carrying the knife in his body, whilst I
retained the gun. I reloaded safely on the rock, thanking God for my
narrow escape.
As the Moor came up with my horse I shouted to him to keep at a
distance, saying the boar was alive and close by. I then got off the
rock and advanced carefully, with both barrels loaded, to the spot to
which I fancied the animal had retreated. Up he got, and came at me
with a rush, receiving the contents of both barrels in his head and
body. I found the long Spanish knife had entered the neck above the
shoulder, and passed along the skin without penetrating the body.
The steel was not good, and had been bent during the struggle. The
boar proved to be a fine three-year-old, with tusks which could have
cut me into shreds. During my tussle with this beast I had a vivid
recollection of having heard that a Moorish hunter, a short time
before my adventure, had fired at a boar at night in a field of Indian
corn, and had followed up the tracks of blood at dawn for some
distance, when he came suddenly upon the wounded animal, who
charged before he could fire, knocked him down and ripped his body
severely. His family, finding next morning he did not return, sent out
in quest of him to the field of corn, and there he was found in a dying
state, wounded in the stomach, just able to relate what had
happened. Within a few yards of the wounded man lay the tusker
quite dead.
Some years ago an English official at Tangier, R———, a very
absent man, sallied out one night to sit for a large boar, which was
reported to pass every evening after dusk a path not far from my
stable at Ravensrock. Near this path in the bush was a rock, on
which my friend squatted with a double-barrelled gun to await the
boar.
It was a very dark night, but the path of white sand in front,
contrasting with the green bush around, could be clearly seen, as
also any object moving along it. He heard the tread of a large animal,
and as it approached within a few feet he fired, but his horror and
dismay can be imagined when down fell a donkey with panniers and
a man on the top! Explanations ensued, with warm expressions of
regret on the part of R———, which were accepted good-naturedly
by the Moor, especially when the former put in his hands double the
value of the donkey and the panniers. The ball had passed through
the top of the skull of the donkey. Strange to say the animal
recovered, and was made use of in R———’s garden.
Boar during the fruit-harvest come down to the orchards near
Tangier and commit great ravages. When sufficient fruit is not
scattered on the ground, they will rub against apple or pear trees
until the fruit falls, or they will spring on the top of a trellis of vines,
tearing it down to the ground to get at the grapes. The Moors put
nooses of rope at the gaps in the hedge where boar enter, and
fasten the noose to a tree or to a bundle of branches. The animal is
often found strangled in the morning; but when the rope is fastened
to loose branches it is less likely to snap, and the boar will carry off
the bundle, until stopped by an entanglement of the rope with some
other object.
Being out one day with a party of hunters, I saw at a distance a
thick bush moving slowly, as by magic, along the top of a dense
copse of gum cistus. No horse or man could be seen. One of the
hunters exclaimed, ‘a boar has been caught in a noose! See the
bush to which it is fastened moving along the top of the copse.’ We
decided to take the animal alive, so approaching the bush and long
rope to which the noose was attached, we laid hold of the rope and
pulled it tight, until the boar was half-strangled. We then gagged the
beast with a thick stick and string. He was dragged out of the thicket,
put on a pack animal and carried to a room in my stable, where the
gag was removed and food and water given.
Next day I invited a party of riders to see the boar turned loose in
the open, two or three miles away from the bush. The horsemen took
no weapons, and our motley pack of boar-dogs were held in leash by
hunters, who were directed to let go when I should give the signal
after the pig had a fair start of one hundred and fifty yards.
Some ladies joined us on horseback, but my wife, being nervous,
rode a donkey, and had a Moor to lead it and to take care of my
young son, who was in front. I placed them on a hillock about two
hundred yards off, where I thought they would be safe and be able to
view the boar. Telling the horsemen and Moors who held the dogs in
leash not to start until I gave the signal, I had the boar conveyed to a
high bank on a dry watercourse, and then removing the gag and
untying the rope, we dropped him gently down, thus giving time for
the men on foot to hide and me to mount before the boar could
charge us. He was only a two-year-old, so his tusks were not very
formidable. The boar bolted up the gulley, and on reaching the top of
the bank looked around, North, South, East and West, but saw no
cover. Viewing my horse about forty yards off he charged, and I
galloped away. The boar halted, looked around, and saw on the
mound an object with brilliant ribbons dangling in the wind, and then
to my great consternation made straight for my wife’s donkey. In vain
I rode full tilt, cracking my hunting-whip, trying to turn the beast, and
shouting to the hunters to let the dogs slip; but before they came up,
the boar got under the donkey, trying to rip it, whilst the Moor,
holding my son aloft on his shoulders, was kicking at the boar.
Up came the dogs, who drew off the boar’s attention, and away
he went; but being better inclined to fight than to gallop, the chase
was short, and he was pulled down by the dogs.
‘Take this knife,’ I said to a long Yankee official; ‘as this is your first
boar-hunt, you shall have the honour of giving the death-blow.’ Knife
in hand, the New Yorker fearlessly advanced, and was inserting
expertly the blade near the region of the heart, when up jumped the
dying pig, knocked over his lank antagonist, and then fell never to
rise again.
Boar when caught young become very attached to man, and will
follow like a dog. They can be taught cleanly habits when kept in a
house, but have no respect for flowers, and cannot resist rooting up
any object which is not firmly fixed in the ground or pavement. I had
a large sow as a pet, which followed me out riding for long distances.
When attacked by dogs on passing villages, the sow would turn
on them and fight gallantly, until I came to the rescue with my
hunting-whip. She became at length very troublesome, and would be
off on the loose into the town whenever the stable-door happened to
be left open. I had frequent complaints from bakers and
greengrocers, and had heavy damages to pay for robberies of bread,
so I gave orders that the sow was to be shut up in a yard.
One day, when the door had been left open, as the sow rushed
rapidly up the street towards a greengrocer’s shop in the little
market-place, where she was accustomed to rob, it happened that a
young mulatto woman, whose legs had been paralysed for some
years, and who gained her livelihood by begging, was crawling on
her elbows and knees along the streets, coming down towards the
Legation. She had never seen a pig in her life, so when she beheld a
large black animal rushing frantically, as she supposed, to devour
her, thought it was a ‘Jin[68].’
The shock was so great, that up she scrambled and ran off; the
paralysis of her legs had ceased. This miracle performed by the sow
was a source of wonder to all, especially to the Mohammedans, loth
to believe that ‘Allah’ should make use of the unclean animal to heal
the maimed. The next day the mulatto appeared at my gate, walking
upright, to petition that I should give her compensation for the fright
she had experienced, pleading also that the pig had deprived her of
the means of gaining her livelihood, for she was now whole, and no
one took pity and bestowed alms on her as before. I gave her only
my blessing, for she was strong and young, and could work. The
sow was presented by me to a gentleman in England, who wished to
introduce a cross of the wild animal.
The sagacity of the boar is greater than that of most animals. A
Moorish Sheikh dwelling in the mountains about forty miles from
Tangier, brought as a gift to the Basha a full-grown boar, that had
been caught when only two months old. The animal had become
very tame; it was brought tied on the back of a pack mule.
A few days after presentation the Basha’s sons carried the boar
out into the country and let it loose, slipping greyhounds to give
chase. The boar knocked over the hounds, charged and ripped two
horses, and got away. Next morning it was found feeding quietly in
the yard of its master’s house, forty miles off! I was glad to learn that
the owner, on hearing how his pet had been treated by the Basha’s
sons, kept the animal until it died.
In the present century lions have rarely been seen in the Northern
province of Morocco.
During a residence of many years I have only heard of two having
been seen in the woods between Tangier and Cape Spartel. I cannot
account for these lions having wandered so far from the Atlas
Mountains—where they are still to be found—except, as the Moors
of those regions relate, that when the winter has been unusually cold
and snow has fallen heavily, the wild animals which dwell in the
higher parts of the Atlas descend to the valleys and plains. Should a
thaw suddenly set in, and rivers and brooks become swollen, the
lions and other wild animals which seek to return to the mountains
are prevented repassing the rapid streams, and stray away from the
district, seeking for forest or for an uninhabited country, and, moving
along the chain of hills to the northward, reach the district of Spartel
—which is about seven miles square—bounded on the western side
by the Atlantic and on the northern by the Straits of Gibraltar.
Early one morning I had a visit from several inhabitants of the
village of ‘Jamah Makra,’ not far from the site of my present villa
‘Ravensrock,’ which stands on a hill, three miles out of Tangier,
surrounded by woods. The men came to request that I should
assemble my hunters and sally out in pursuit of a wild animal which,
they related, had lacerated with its claws the flank of a mare and
bitten it in the neck. They informed me that they had been roused in
the middle of the night by the tramp of horses galloping through the
lanes—snorting and neighing—and supposed that cavalry had been
sent to surround the village. But to their surprise they found their own
ponies (which are allowed to run loose on the hills when not required
for agricultural purposes, and live in a half-wild state, never allowing
man to approach them, especially at night-time) had by instinct
sought safety in the village, trying to penetrate even into the huts.
Amongst the herd was the wounded mare, in a dying state.
I assembled a party of hunters with their boar-dogs, and
proceeding to the spot we found round the village tracks of a large
animal; evidently of the feline race, as the footprints were round, with
no mark of nails, but had pads, as in the print of a cat’s foot. The
beast appeared to have avoided as much as possible the open path,
and to have walked near or amongst the ilex bushes, on which we
found long tawny hairs, showing it was a male lion. We also came
across the half-eaten carcasses of a boar and of a porcupine. There
were marks too as of a herd of boar making a stampede in a
southerly direction, fleeing from the dread monarch of the woods.
We turned our dogs into the thicket—where, by the tracks, we
knew the lion had entered—and placed two guns at each run. But
the dogs returned from the thicket and shrank behind their masters.
They had evidently come upon or winded the lion, and we could not
induce them to hunt. The beaters, after entering the thicket, firing
guns, and beating drums, refused to advance further; so we had to
abandon the hunt.
A woman whom we met informed us that, on going to a fountain in
her orchard to draw water, she had met a ‘jin’ (evil spirit), evidently,
from her description, a lion; that she became paralysed from fright
and could not move; that the ‘jin’ had eyes like lamps, and after
gazing at her had turned aside into the bush.
The Moors believe that lions will never attack a nude woman,
such is the magnanimous beast’s delicate sense of shame.
Lionesses, it is to be concluded, are less particular. The dame did
not mention that she had a knowledge of this, so we know not
whether she dropped her vestments to save her life.
There was a good moon; so I determined to sit for the lion, safely
perched on a rock, where, though it would be possible for a lion to
climb, yet I should have had a great advantage in an encounter with
gun and pistols. I passed the night in a state of excitement—starting
at every rustle made by rabbit, ichneumon, or even rats—without
seeing anything of the king of beasts. But about midnight I heard
what sounded from a distance like the deep bellow of a bull.
A few days later, hearing that the track of the lion had been seen
at ‘Ain Diab,’ a wood near Cape Spartel, I collected the hunters and
rode to the ground, about eight miles from Tangier. There we tracked
the lion into a dense thicket. The dogs again refused to hunt, as on
the previous occasion, winding no doubt the lion. This was good
proof that he was at home; so posting the guns, I directed the
beaters to drive the wood from the foot of the hill and that guns
should occasionally be fired and drums beaten.
A few minutes after I had taken up my post a Moor hurried up to
where I was standing, in a great state of excitement, pale as death,
saying, ‘I have seen the man[69]!’ ‘What man?’ I asked. He repeated,
‘I have seen the man! I had entered the thicket to look at an olive-
tree from which I thought I could cut a good ramrod; there is a rock
rising about twenty feet above the olive-tree, and as I stooped to look
whence I could best cut a branch, I saw a great shaggy head, with
fierce eyes glaring at me from between two huge paws. I had laid
down my gun to cut the olive stick; I dared not turn to take it up
again, so left it there and crawled back through the bush to tell you
what I have seen.’
The rock, which he then pointed out, was about two hundred
yards from where we stood. I collected the sportsmen and selected
three of them (my brother and two Moors upon whose courage I
could depend), and we determined to beard the lion in his den. My
left arm was in a sling, having been injured while playing cricket a
few days previously. As we advanced into the dense thicket I was
prevented, by the pain caused by the branches knocking against my
arm, from following quickly my companions. Carried away by their
desire to slay the lion, they rushed on headlong, regardless of wait-
a-bit thorns and other impediments; so I was left in the lurch. Feeling
uncertain about the exact direction they had taken, but hearing, as I
thought, the sound of some one passing in front of me, I shouted,
‘Where are you? why are you returning?’ No reply. Yet it was evident
the moving object had approached me within a very few yards. Again
I called, ‘Why don’t you speak?’ Then I heard a rush, as I suddenly
came to an open spot of sandy soil, upon which I could trace the
footmarks of the lion who had just passed. The animal had evidently
moved away from the rock when he heard or saw the three men
approaching, and having no desire to attack man unprovoked, had
doubled back, passing close to me. All this flashed through my brain;
I halted, kept perfectly still, holding my breath, for I had not the
courage, alone and with an injured arm, to follow the dread beast.
Moreover, I could never have caught it up, at least I tried so to
convince myself, and thus to hush any feeling of shame at my
cowardice.
My companions returned a few minutes afterwards, reporting that
they had reached the rock where the lion had been; but he had
evidently left on their approach, and they had tracked him through
the bush to the spot where I had stood when he passed. We
followed the direction the lion took for some time without success,
and we supposed he must have made off at a swinging trot.
The following day we heard that an ox had been killed on the hills
of Anjera between Tangier and Tetuan, and that the lion had gone in
the direction of the snow-topped mountains of Beni Hassén.
On each visit of a lion to the Tangier district the track of a hyena
had been seen to follow that of the sultan of the forest.
On one occasion, when there were rumours of a lion having been
heard of in the Tangier district, and we were out hunting boar in the
woods near Spartel, I heard several shots fired from the side of a hill
where I had posted the guns, and a beater shouting to me, as I stood
hidden behind a small rock in some low bush, ‘“Ya el Awar!”—Oh ye
blind! The lion to you!’ An instant after I viewed, bounding over the
bushes, a large shaggy animal. With its huge mouth open and
bristling mane, it looked very terrible; but I knew at once it was not a
lion; so I waited till the beast was within a few yards and sent a bullet
through its heart. It turned out to be a very large Hyena rufus—
striped, not spotted—larger than any specimen of that animal I have
seen in the Zoological Gardens or any menagerie.
The stench of the animal was overpowering; the skin was in
beautiful condition, and proved very handsome when preserved.
A grand lion was seen many years ago, standing in the early
morning on the sand-hills which line the beach close to the town of
Tangier, and causing great alarm. But it turned out to be a tame lion
which a ‘Shloh’ woman—who, as a Sherífa, was endowed with a
slight halo of sanctity—had brought captive from the Atlas
Mountains. She led it about with only a loose rope round its neck, as
she begged from village to village, and had arrived outside the gates
of Tangier the previous evening, after they were closed, and she had
laid down to sleep near the lion, which, during the night, had strayed
away. This lion was quite tame and harmless, and came back to her
from the sand-hills when she called it.
A Spanish gentleman told me that returning home late one dark
night from a party in Tangier, carrying a small lantern to light his way,
he saw what he fancied was a donkey coming towards him in one of
the very narrow streets of the town where two stout persons on
meeting can hardly pass each other. He turned his lantern on the
object, and, to his dismay, saw the glistening eyes and shaggy head
of a lion which he had already seen led in daytime by the woman
through the streets. The beast was alone, without its keeper. The
Don said he had never made himself so small as when he stood
against a closed door to allow his Majesty to pass; which he did quite
pacifically.
‘Oh ye blind! The lion to you!’
This accusation of blindness is perhaps the mildest form of abuse
employed by the beaters, in the excitement of the hunt, to the guns
posted to await the boar. Sir John, as Master of the Hunt, shared in
the very liberal abuse indulged in by the men who had laboriously
driven the boar from the thick coverts towards him and his friends,
native and foreign, who waited to shoot the pigs as they broke. Every
possible term of abuse—and Arabic is rich in such—together with
imprecations such as only Oriental imagination could devise, would
be yelled at them as a warning not to miss. Strangers too would
always be indicated by any peculiarity in their appearance or dress.
Neither did the excited beaters, at such moments, put any check on
their rough wit. But the railing of Moorish sportsmen at each other,
however violent in the ardour of the chase, is never resented.
As a case in point, Sir John related the following story.
A former Governor of Tangier, a thorough sportsman, was out
hunting on one occasion, when a man of low degree who was acting
as beater, and, as is usually the case, had his own dogs with him,
started a boar in the direction of the Basha, who was sitting near the
animal’s expected path ready to receive him. The beater called out,
swearing lustily at the Basha, and using every opprobrious term he
could think of; adding that if he missed his shot he should never be
allowed to fire again!
The Basha fired and killed the boar.
Some little time after, when the beat was finished, the huntsmen
assembled as usual, and the Basha asked who it was that had
started the boar he had shot. The poor beater, feeling he had
exercised the licence of the chase rather too boldly, kept somewhat
in the background, but, on this challenge, came forward and
acknowledged that it was he who had done so.
‘And what did you shout out to me when the boar took in my
direction?’ asked the Basha. The beater, dismayed, was silent. But
on the question being repeated, acknowledged having called out,
‘The boar to you—oh blind one!’
‘Only that!’ exclaimed the Basha. ‘Surely I heard you abuse me.
Tell me what you said.’
In reply to this the beater, in desperation, burst out with all the
abuse he had uttered. Whereupon the Basha, taking from his wallet
four ‘metskal’ (then worth some three Spanish dollars), presented
them to the beater, saying, ‘Take this. I know you were anxious on
account of your dogs, and for the success of the sport. I pardon your
abuse of me.’

After his retirement from his official position, Sir John lived little
more than seven years, dividing his time between Morocco and
Europe, returning, as has been said, for the winter to his beloved
‘Ravensrock,’ enjoying his sport to the end, and at intervals jotting
down his ‘Scraps from my Note-book’ as a slight record of his life. ‘I
feel,’ he says, referring to the appearance of some of his stories in
Murray’s Magazine, ‘like a dwarf amongst tall men. Never mind. If
my relatives and friends are pleased and amused, I shall continue to
unwind the skein of my life till I reach my infancy.’ Among the last of
the notes made by Sir John in his ‘Note-book’ was the following,
which may be appropriately introduced at the close of this sketch of
his career.

Body and Soul.

‘The death of the aged is always easy,’ said the F’ki Ben Yahia,
‘compared with the death of the young.’
‘This arises,’ continued the F’ki, ‘from the willingness with which
the immortal soul is glad to flee from an aged body, corrupted by a
long residence in this world, and from disgust at the sin and
wickedness into which it has been plunged by the depravity of the
body. Whereas, the young body and soul are loth to part; for the soul
rejoices in the innocent enjoyments of youth and the harmless
pleasures of this world, and to separate them is, as it were, to
separate the young damsel from her first pure love.’
‘Oh, merciful God!’ exclaimed the F’ki, ‘put away the corruption of
my body, and teach me to follow the purer inspiration of the soul
which was breathed into me by Thee, O Almighty and
Incomprehensible God!’

In Berwickshire, at Wedderburn Castle, a place then rented by

him, Sir John Hay Drummond Hay died on the evening of Monday,
Nov. 27, 1893.
He was buried in the churchyard of Christ Church, Duns. A few
days after the funeral one of the family received a letter from a
member of the British Legation at Tangier, in which he mentioned
that on going to the Legation on the morning of Nov. 27, he was
surprised to see the British flag at half-mast, and, calling to the
kavass in charge, reprimanded him for his carelessness, directing
him to take the flag down.
The kavass excused himself, saying that, while hauling down the
flag the previous evening, the halyard had broken, and he had
consequently been unable to lower the flag further; but that he had
sent for a man to swarm the mast and repair the halyard and thus
release the flag. This however, the writer added, was not
accomplished till next morning.
Thus it happened that while the man was passing away who for
forty years had represented Great Britain in Morocco, the British flag
remained at half-mast.
 INDEX.

Abbas Pasha, 26.

Abbotsford, 3.
Abd-el-Hadi, 136.
Abd-el-Kader, 69, 71, 72.
Abd-el-Kerim, 227.
Aberdalgie, 6.
Aberdeen, Lord, 44, 66, 68.
Acre, 30.
Addington, H. C., 135; his letter to Sir J. D. Hay, 140.
Agadir, 317.
Agraz battle, 160.
Ahal Kubla tribe, 297.
Ahmed Ben Ali, 381.
Ain Dalia, 79.
Ain Diab wood, 392.
Ain-Umast fountain, 122.
Aisa, 228.
Aisawa, or snake-charmers, 177.
Aji, 236.
Akba el Hamra hills, 85.
Akhlij village, 292.
Alarbi el Saidi, 184.
Alcalá, 14.
Alexandria, 24.
Alfred, Prince, at Tangier, 203.
Alhádari, 147.
Alhucema, Island of, 148.
Ali Bufra, 241.
Alison, Charles, 49, 73.
Allen, Mr., 347.
Amar, Sheikh, 252.
Andersen, Hans Christian, ‘In Spain,’ 222, 225; his description of the old
Legation, 222; his letter to Lady Hay, 225.
Anjera, 186, 393.
Arab dance, 91.
Arára, 366, 374.
Argan tree, 122.
‘Arum arisarum’ or yerna, 325 note.
Ashkar, caves of, 364.
Assuad, Sultan, 99; inscription on his tomb, 99 note.
Athol, H.M.S., 6.
Atlas Mountains, 116, 272, 276, 289, 293, 390; valley, 290.
Austria, Emperor of, 363.
Awára plain, 185, 369, 371, 376.
Awínats wood, 376.
Azaila, 137, 242.
Azamor, 169.
Azdot, 142.

Bab-el-Haddad, or the Smithy Gate, 99.

Bab-el-Khemés, or the Thursday-gate, 111.
Bab-el-Mahsen, or the Government-gate, 289.
Bab Hamár gate, 107.
Bab Khadár, or ‘the Green Gate,’ 112.
Bakáli, Sheríf, 283.
Ball, Mr. J., 88 note.
Bankhead, Charles, 49.
Bardlaiimi, 129.
Barker, Mr. Burchardt, 16.
Barnett, Colonel, 63.
Barseset, Izak, 128.
Basha Hamed, 223.
Beehive, a Moorish, 195 note.
Beheira u el Gintsor, 269.
Bell, Dr., 23.
Ben Abd-el-Sadek, 364.
Ben Dawud, 275.
Ben Dris, the Grand Uzir, 113, 301; conferences with Sir J. D. Hay, 113, 115,
117.
Ben Isa, 376.
Ben Nasr, F’ki Sid Mohammed and Zarhoni, dialogue between, 81.
Ben Nis, 283.
Benabu, Governor of Tangier, 184; story of his arrest and death, 184-192; his
system of governing, 193.
Beni Aros, 210.
Beni Gorfet mountain, 242.
Beni Hassén mountain, 393; tribe, 90.
Beni M’suar mountain, 242, 371.
Benibugaffer village, 144, 146.
Benisargan, Jakob, 128.
Benshiten, 131.
Berbers, origin of the, 291.
Besika Bay, 60.
Beyrout, 30.
Birra Burub pass, 269.
Boar-hunting, 366-389.
Bojador, Cape, 317.
Bokhari guards, 119, 274.
Bonelli eagle, 367.
Borj Ustrak, 158.
Bosco, his sleight of hand, 33.
Briant, 43.
Brij, 368.
Brooks, Mr., 296.
— Mrs., 363.
Broussa, 51.
Bu Amar hill, 371.
Bubána river, 185.
Buceta, Colonel, Governor of Melilla, 149.
Buchanan, Mr., 208.
Bugeaud, Maréchal, 79.
Bulwer, Sir Henry, 68.
—, Mr., 67, 68, 70, 71, 185.
Buyukdere, 50.

Cadiz, 225.
Campbell, Colonel, 28.
Canning, Lady, 47, 66.
Canning, Sir Stratford, 47, 58, 66; appointed Ambassador at Constantinople,
49; his method of conducting business, 49; letter from Sir J. D. Hay on the
state of affairs in Tangier, 68-71.
Carstensen, Mr., 142, 223.
Cartwright, Mr., 66.
Castelar, Señor, 207.
Cattle-lifting in Morocco, 193.
Ceuta, its advantages over Gibraltar, 234.
Chapman, Mrs., anecdote of Sir J. D. Hay, 164.